Estimating Earnings Adjustment Frictions : Method and Evidence from the Social Security Earnings Test

Recent literature has documented that individuals face frictions in adjusting their earnings with respect to policy, but this literature has not yet developed a method for estimating earnings adjustment costs. We introduce a method for estimating the cost of adjusting earnings, as well as the elasticity of earnings with respect to the incentive to earn. Our method uses information on the amount of bunching in the earnings distribution at convex budget set kinks before and after policy changes in the earnings incentives around the kinks: the larger is the adjustment cost, the smaller is the absolute change in bunching from before to after the policy change. We apply this method in the context of the convex kink created by the Social Security Annual Earnings Test (AET). Using a one percent sample of earnings histories from Social Security Administration micro-data, we show that individuals continue to bunch at the kink formerly created by the AET even when they are no longer subject to the AET, indicating that they must face adjustment frictions. We estimate in a baseline case that the earnings elasticity with respect to the implicit net-of-tax share is 0.35, and the Öxed cost of adjustment is around $280. Our results demonstrate that the short-run impact of changes in the e§ective marginal tax rate can be substantially attenuated, which can inform projections of the timing of earnings responses to tax and transfer policies